Latest News

Daily AI briefings grounded in primary sources

Today

June 7, 20264 briefings today

NVIDIA Nemotron 3 Ultra processor visualization with open-weight model architecture and agentic ecosystem connections in deep space

NVIDIA's Nemotron 3 Ultra: Open 550B MoE Built for Long-Running Agents

NVIDIA ships Nemotron 3 Ultra: a fully open 550B MoE model with 1M-token context, 5× faster inference, and Day-0 LangChain coalition backing for agents.

June 7, 20262 min read

ToolsNotable

LangChain's Platform Blitz: Managed Deep Agents and LangSmith Pillars

LangChain shipped Fleet, Engine, Sandboxes GA, and Managed Deep Agents in one week — a coordinated platform push backed by Harmonic's 4× retention case study.

June 7, 20262 min read

IndustrySignificant

Microsoft Launches Seven MAI Models in OpenAI Independence Play

Microsoft unveiled seven in-house MAI models trained from scratch, led by MAI-Thinking-1 scoring 97% AIME — a direct bid to reduce OpenAI dependency.

June 7, 20262 min read

ResearchMajor

Anthropic RSI Report: Claude Writes 80% of Its Own Code, 52× Training Speedup

Anthropic discloses Claude now writes 80% of its own codebase and delivered a 52× training speedup — with human judgment the last narrowing frontier.

June 7, 20262 min read

Earlier briefings, by category

Industry

39 articles

View all in Industry

IndustryMajor

Anthropic Confidentially Files S-1 at $965B Valuation

Anthropic filed a confidential S-1 at a $965B valuation and $47B revenue run-rate, joining OpenAI and SpaceX in the 2026 frontier AI IPO wave.

June 6, 20262 min read

IndustryNotable

Enterprise AI ROI Reckoning: Token Spend Meets Hard Returns

Uber burns its 2026 AI budget in 4 months, hyperscalers post negative implied returns, and H200 rents crater — enterprise AI ROI accountability is here.

May 30, 20262 min read

IndustrySignificant

Anthropic's $65B Series H Crowns It the World's Most Valuable AI Startup

Anthropic closes a $65B Series H at $965B valuation — surpassing OpenAI — as run-rate revenue hits $47B and Claude captures 54% of the corporate coding market.

May 30, 20262 min read

IndustryNotable

AI Infra Capital Surge: Four Deals, $13B+ in One Day

On May 26, four capital events — Baseten, OpenRouter, Suno, and Micron — signal sustained market conviction across the full AI infrastructure stack.

May 29, 20262 min read

IndustrySignificant

Starbucks Kills NomadGo; Hyperscaler AI ROI Turns Negative

Starbucks shut down NomadGo nationwide in May 2026 after inventory AI failures. FT data shows hyperscaler ROI turned negative: Google −15%, Meta −28%.

May 29, 20262 min read

IndustrySignificant

Microsoft, Uber Confront Real AI Costs as Subsidy Era Ends

Microsoft dropped Claude Code licenses over token costs, Uber burned its 2026 AI budget in four months — the AI subsidy era is ending for enterprise customers.

May 27, 20262 min read

IndustrySignificant

Karpathy at Anthropic: Claude Accelerates Its Own Pretraining

Karpathy joins Anthropic's Pretraining team to use Claude for accelerating its own research — a landmark hire confirmed across three sources.

May 23, 20262 min read

IndustryNotable

Andrej Karpathy Joins Anthropic for Frontier LLM R&D

Andrej Karpathy confirms joining Anthropic for frontier LLM R&D, calling the next few years 'especially formative' — and open-source implications loom large.

May 19, 20262 min read

IndustryNotable

Anthropic Doubles Claude Limits via SpaceX's 220K-GPU Colossus Deal

Anthropic has doubled Claude token limits across all paid plans and lifted peak-hour caps, powered by a new SpaceX deal giving access to 220K+ Nvidia GPUs.

May 19, 20262 min read

IndustryNotable

Anthropic in $30B+ Talks at $900B as Cerebras IPO Pops 68%

Anthropic is in early talks for a $30B+ raise at a $900B valuation as Cerebras popped 68% on its Nasdaq debut — AI capital markets hit two structural milestones in one week.

May 17, 20262 min read

IndustrySignificant

Anthropic Passes OpenAI in Enterprise; Coding Agent War Ignites

Menlo Ventures data shows Anthropic at 34.4% enterprise share vs OpenAI's 32.3% in April, triggering simultaneous free-trial counter-offers from both labs.

May 15, 20262 min read

IndustrySignificant

OpenAI Launches $4B DeployCo to Solve the Enterprise AI Last Mile

OpenAI's $4B Deployment Company embeds forward-deployed engineers inside enterprises, directly challenging consultancies on the AI implementation last mile.

May 12, 20262 min read

IndustrySignificant

Coinbase Cuts 700 Citing AI — DOL Finds 13,000 IT Jobs Gone in April

Coinbase's AI-attributed 14% RIF is a mid-cap public-tech first. Dorsey, PayPal, and April DOL data showing 13K IT jobs shed confirm this is structural.

May 10, 20262 min read

IndustryNotable

Anthropic Closes in on $1 Trillion With $50B Raise, October IPO

Anthropic is negotiating up to $50B at a $900B–$1T pre-money valuation, with Dragoneer, General Catalyst, and Lightspeed in talks and an October 2026 IPO targeted.

May 9, 20262 min read

IndustrySignificant

Murati Testifies Altman Bypassed OpenAI Safety Board Under Oath

Mira Murati testified under oath that Altman misled her about safety board clearance; Brockman's journal and firing-weekend texts now key trial evidence.

May 9, 20262 min read

IndustryNotable

Anthropic Secures All of Colossus 1 — Rate Limits Doubled Same Day

Anthropic secures all of SpaceX's Colossus 1 (220K GPUs, 300 MW) and doubles Claude Code rate limits—sub-agent fan-out is now viable at $200/month.

May 8, 20262 min read

IndustryNotable

OpenAI Trial: Sworn Testimony Pins Altman Firing on Trustworthiness

Murati's sworn testimony in Musk v. Altman confirms the 2023 board firing was about trustworthiness, while Brockman testifies to 1,667× compute growth.

May 7, 20262 min read

IndustrySignificant

Anthropic Lands Colossus 1, Ships Managed Agents in One Day

Anthropic leases all of SpaceX Colossus 1 (300 MW, 220K GPUs), doubles Claude Code limits, and ships Managed Agents, Dreaming, and Outcomes in one day.

May 7, 20262 min read

IndustrySignificant

Anthropic Targets Wall Street: 10 FSI Agents and $1.5B Goldman JV

Anthropic ships 10 Claude agents for financial services and forms a $1.5B JV with Goldman Sachs and Blackstone as Wall Street's new operating layer.

May 7, 20262 min read

IndustryNotable

OpenAI Ships GPT-5.5 Instant as ChatGPT's Default Model

OpenAI's GPT-5.5 Instant becomes the default ChatGPT model, with factuality gains, memory personalization, and 86% free-tier GPQA benchmark score.

May 6, 20262 min read

IndustryNotable

OpenAI Co-Founders Held Undisclosed Cerebras Equity, Trial Reveals

Brockman confirmed under oath that he held undisclosed Cerebras equity while OpenAI committed $20B+ to the firm. California law calls this self-dealing.

May 5, 20262 min read

IndustryNotable

$16.20 vs $2.20: The ARPU Gap That Explains Anthropic's Lead

Anthropic crossed $30B ARR on 134M users—$16.20 per user monthly vs OpenAI's $2.20 on 900M. Claude Code's enterprise stickiness is driving the structural gap.

May 5, 20262 min read

IndustrySignificant

GPT-5.5 Breaks OpenAI's Revenue Records — Codex Doubles in One Week

GPT-5.5 posted OpenAI's strongest-ever launch: API revenue 2× faster than any prior release, with Codex doubling revenue in 7 days on enterprise agentic coding demand.

May 3, 20262 min read

IndustrySignificant

DeepSeek V4 Pro: Open-Source Frontier, 1M Context, 50–100× Cheaper

NIST CAISI confirms DeepSeek V4 Pro as top Chinese model, ~8 months behind US leaders — but MIT-licensed, 1M context, and 50–100× cheaper than closed rivals.

May 3, 20262 min read

IndustryNotable

Agents for Everything Else: Coding Tools Break Into Knowledge Work

swyx's AI Engineer London keynote and Karpathy's Sequoia chat both chart the same 2026 shift: coding agents escaping the dev stack into all knowledge work.

May 1, 20262 min read

IndustryNotable

OpenAI Codex Reaches 3M Users With Full Agent OS Architecture

Codex hit 3M weekly users — tripling since January — as OpenAI revealed a full agent OS: plugins, sub-agents, automations, hooks, and Guardian Approvals.

April 30, 20262 min read

IndustryNotable

Anthropic Plays All Three Boards: Design Tool, Embedded Engine, Managed Agents

Anthropic reveals three Claude shapes simultaneously: standalone design surface, embedded model defaults in enterprise stacks, and managed agent infrastructure.

April 30, 20262 min read

IndustryNotable

OpenAI's $852B Valuation Faces a CFO Mutiny and a Musk Trial

WSJ reveals OpenAI's CFO warned on compute funding, then was excluded from key infrastructure meetings. Musk trial opens as $852B valuation comes under pressure.

April 29, 20262 min read

IndustryNotable

OpenAI Ends Microsoft Exclusivity — Amazon Inherits the Cloud Route

OpenAI goes multi-cloud as Microsoft exclusivity ends; Amazon becomes the primary delivery route. WSJ reveals missed targets, CFO warns on compute funding.

April 28, 20262 min read

IndustryNotable

67% of US AI Data Centers Are Heading Rural — and Towns Are Resisting

Pew data: 67% of planned US data centers target rural areas. Archbald, PA shows how small towns are pushing back against AI infrastructure expansion.

April 28, 20262 min read

IndustrySignificant

OpenAI's Agent Stack in Focus: Workspace Agents, Symphony, Codex

Three OpenAI signals converge on one enterprise-workflow OS: Workspace Agents, Symphony, and Codex — all accelerating at once. Free window closes May 6.

April 28, 20262 min read

IndustrySignificant

DeepSeek V4: 1M-Context Open Weights, 1/7 Opus 4.7 Pricing

DeepSeek V4 drops two open-weight models with 1M-context by default, CSA+HCA hybrid attention, and V4-Pro priced at roughly 1/7 Opus 4.7's output cost.

April 26, 20262 min read

IndustryMajor

GPT-5.5: Agentic-First Model, 82% Terminal-Bench, Safety at HIGH

OpenAI's GPT-5.5 arrives six weeks after 5.4 with a 7-point Terminal-Bench gain, doubled pricing, and cyber/bio safety classifications at HIGH.

April 26, 20262 min read

Industry

SpaceX's $60B Cursor Option: Frontier-Lab API Dependency as Weaponizable Risk

SpaceX holds a $60B call option on Cursor with a $10B breakup floor — structured as much against frontier-lab API dependency as toward acquisition.

April 24, 20262 min read

Industry

Google Reframes Vertex AI as Gemini Enterprise Agent Platform at Cloud Next

Google rebrands Vertex AI as the Gemini Enterprise Agent Platform at Cloud Next, adding 200+ models and five major consulting partnerships to close the enterprise adoption gap.

April 23, 20262 min read

Industry

OpenAI Offers Free ChatGPT to US Clinicians, Releases HealthBench

OpenAI makes ChatGPT free for verified US medical professionals and releases HealthBench Professional — an open benchmark that GPT-5.4 outperforms on against physicians.

April 23, 20262 min read

Industry

OpenAI Launches ChatGPT Workspace Agents for Enterprise Plans

OpenAI launches ChatGPT Workspace Agents in research preview — cross-tool automation agents for Business, Enterprise, Edu, and Teachers plans.

April 23, 20262 min read

Industry

Anthropic's Compute Shortfall Has Become a Trust Crisis

Dario Amodei's conservative 2024 capex call is now triggering opaque quota changes, harness bans, and tokenizer inflation — handing OpenAI a sustained PR windfall.

April 23, 20262 min read

Industry

OpenAI Launches GPT-5.1 with Native Agent Infrastructure

OpenAI released GPT-5.1 with built-in agent orchestration, persistent memory, and native tool invocation — positioning against Anthropic and Google in the agentic AI infrastructure race.

April 14, 20263 min read

Technology

30 articles

View all in Technology

TechnologyMajor

Microsoft's 7 MAI Models and MAIA 200 Chip Signal OpenAI Exit

Microsoft Build 2026: seven MAI models, MAIA 200 chip (30% better than Nvidia GB200), and Scout agent in Teams mark a chip-to-harness self-sufficiency push.

June 6, 20262 min read

TechnologySignificant

Claude Opus 4.8 Ships with Dynamic Workflows and 2.5x Fast Mode

Anthropic ships Claude Opus 4.8 with dynamic workflows: hundreds of parallel subagents plus adversarial judges — and a honesty-first architecture.

May 30, 20262 min read

TechnologyNotable

Gemini Embedding 2: Google DeepMind's First Native Multimodal Embedder

Google DeepMind publishes the GE 2 white paper — its first embedding model natively unifying text, audio, video, and image in one shared vector space.

May 29, 20262 min read

TechnologySignificant

Anthropic Launches Opus 4.8 with Parallel Sub-Agent Workflows

Claude Opus 4.8 ships with Dynamic Workflows that spawn hundreds of parallel sub-agents and a 3× cheaper Fast mode, at the same standard price as 4.7.

May 29, 20262 min read

TechnologySignificant

Google I/O 2026: Antigravity Emerges as Google's Managed Agent Platform

Google I/O 2026 deprecated Gemini CLI in favor of Antigravity, launched Gemini for Science with 100+ institutions, and unveiled native video editing via Gemini Omni Flash.

May 27, 20262 min read

TechnologySignificant

SynthID Coalition Locks In: OpenAI, ElevenLabs, and Kakao Join Google's Watermark Standard

Google's SynthID watermark is becoming the cross-industry AI content provenance standard, with OpenAI, ElevenLabs, and Kakao joining NVIDIA in the coalition.

May 23, 20262 min read

TechnologyMajor

Google I/O 2026: Full-Stack AI Overhaul from Chips to Agents

Google I/O 2026 drops Gemini 3.5 Flash, Antigravity 2.0, Gemini Spark, XR glasses, and 8th-gen TPU — the most comprehensive AI product push in the company's history.

May 23, 20262 min read

TechnologySignificant

Google I/O 2026: MCP, A2A, and AGUI Settle as the Core Agentic Stack

Google I/O 2026 confirms three agent protocols as the settled core stack—MCP, A2A, AGUI—while A2UI, AP2, and X42 remain contested. The real risk is the operating surface.

May 19, 20262 min read

TechnologyMajor

2026's Defining AI Lever: Harness Architecture, Not Model Choice

A 32,000-GPU-hour benchmark confirms the harness layer outweighs model choice — identical backbones swing 3× in accuracy depending on agent framework.

May 17, 20262 min read

TechnologySignificant

Shai-Hulud Worm Poisons 373 npm Packages via CI Cache, Jumps to PyPI

The shai-hulud worm exploited TanStack's CI cache to poison 373 npm package versions across 169 packages — including Mistral AI — before jumping to PyPI.

May 15, 20262 min read

TechnologySignificant

Karpathy and Thariq Make the Case for HTML as AI's New Output Standard

Anthropic engineer Thariq argues HTML beats Markdown for AI output; Karpathy backs it with a six-step format evolution toward interactive neural video.

May 12, 20262 min read

TechnologyNotable

GPT-5.5 Instant Rolls Out with 52.5% Fewer Hallucinations

OpenAI rolls out GPT-5.5 Instant to ChatGPT and the API, claiming 52.5% fewer hallucinations on high-stakes prompts in medicine, law, and finance.

May 10, 20262 min read

TechnologyNotable

Claude Mythos: 16-Hour METR Horizon, 3-Week Palo Alto Validation

METR puts Claude Mythos at a 16-hour task horizon, 2× the next best. Palo Alto Networks: 3 weeks AI-assisted equaled a year of manual penetration testing.

May 9, 20262 min read

TechnologyNotable

OpenAI Ships Three Realtime Voice Models with GPT-5 Reasoning

OpenAI launched three voice models at once in its Realtime API: GPT-Realtime-2 brings GPT-5 reasoning; Translate and Whisper add streaming capabilities.

May 8, 20262 min read

TechnologySignificant

Claude Mythos Validated: Firefox Fixes 15 Months of Bugs in One April

Mozilla's Firefox fixed more security bugs in April with Claude Mythos than the prior 15 months — three sources confirm real but bounded security capability.

May 8, 20262 min read

TechnologyNotable

Sub Quadratic's subQ: Sparse Attention Claims Under Scrutiny

subQ claims 12M-token context at 52× FlashAttention speed, but benchmarks test only the 1M preview model, with figures differing between video and website.

May 6, 20262 min read

TechnologyNotable

AI Agent Finds Universal Linux Privilege-Escalation in One Hour

Theori's AI agent found CVE-2026-31431 in 1 hour — a universal Linux LPE dormant since 2017. CISA added it to KEV; CrowdStrike confirms active exploitation.

May 5, 20262 min read

TechnologySignificant

Agent Harness Engineering: Same Model, 6x Performance Variance

Four sources — Tsinghua papers, Melbourne ICL study, AgentFloor benchmark, deepagents-cli — converge: harness design drives a 6x model performance spread.

May 5, 20262 min read

TechnologyNotable

Patrick Debois Formalizes Context Engineering's CI/CD Moment

Patrick Debois, who coined 'DevOps' in 2009, introduces the CDLC — a 4-phase framework applying CI/CD rigor to AI agent context engineering.

May 4, 20262 min read

TechnologyNotable

Claude Opus 4.7 Tops Coding Benchmark and Powers Six Hackathon Winners

A peer-reviewed AlphaZero benchmark and a global hackathon both confirm Claude Opus 4.7 as the current frontier in agentic coding.

April 30, 20262 min read

TechnologySignificant

GPT-5.5's Pre-Train Lift Resets the Frontier Benchmark

GPT-5.5 scores 87.3 vs Opus 4.7's 67.0 on 23 exec deliverables — a pre-train jump, not inference tricks — and is the first frontier model to catch planted fake migration data.

April 29, 20262 min read

TechnologyNotable

Gemma 4 Beats Models 20× Its Size on LM Arena, Goes Apache 2.0

Google DeepMind's Gemma 4 family launches under Apache 2.0 with MoE architecture, on-device multimodal, and the 31B dense model ranked #3 on LM Arena.

April 28, 20262 min read

TechnologyNotable

DeepSeek-V4 and Kimi-K2.6 Shift the Open-Weights Agentic Baseline

DeepSeek-V4's MIT-licensed 1M-context MoE and Kimi-K2.6's multimodal orchestration create the first complete open-weights agentic deployment stack.

April 27, 20262 min read

TechnologySignificant

GPT-5.5 in Codex: Builder Euphoria, Skeptic Alarm, Toolchain Rush

Three independent sources captured GPT-5.5 from every angle simultaneously: builder euphoria, toolchain adoption, and a structural reliability alarm.

April 27, 20262 min read

TechnologyNotable

Kimi K2.6 Becomes Open-Source #1 with 300-Agent Swarms

Moonshot AI's Kimi K2.6 leads the open-source index with 300 concurrent sub-agents, 4,000 tool calls, and a 12-hour autonomous coding marathon.

April 26, 20262 min read

TechnologySignificant

Google Ships Deep Research Max with MCP and $4.80 Per-Report Pricing

Google Deep Research Max costs $4.80/report and uses MCP to connect to private data stores. Independent 7-task testing shows the cheaper tier wins 5 of 7.

April 26, 20262 min read

Technology

GPT Image 2 Wins 93% of Blind Tests — Reasoning Joined the Visual Stack

GPT Image 2 claims a 26-point lead in Image Arena blind tests — unprecedented for the category — by wiring a reasoning loop before every pixel render.

April 25, 20262 min read

Technology

DeepSeek V4-Pro: 10× KV Cache Efficiency at Open-Source Scale

DeepSeek V4-Pro launches with 1.6T parameters, 1M context, and 10× KV cache reduction over V3.2 — multiplying inference concurrency roughly 10× on the same hardware.

April 24, 20262 min read

Technology

GPT-5.5 Reframes AI Progress as Intelligence Per Token

GPT-5.5 scores 2.5× better intelligence-per-token than 5.4, surpasses the human baseline on OS World, and expands Codex into a full desktop agent.

April 24, 20262 min read

Technology

Qwen3.6-27B Surpasses a 397B Model on Coding Benchmarks

Alibaba's Apache 2.0 27B model outperforms Qwen3.5-397B-A17B on all major coding tasks and runs locally on 18 GB RAM — 'bye bye subscription era' claims are spreading.

April 23, 20262 min read

Strategy

10 articles

View all in Strategy

StrategyMajor

The AI ROI Reckoning: IBM, Galloway, Marcus Sound the Same Alarm

Five independent reports agree: 95% of enterprise AI spend has no nameable ROI, IBM says $6–8T capex has no revenue to match, and a 50–70% correction looms.

June 6, 20262 min read

StrategyNotable

Anthropic's '2028' Essay Draws Battle Lines on US-China AI and Open Source

Anthropic's '2028' essay frames the AI finish line as recursive self-improvement, splitting the industry on compute restriction vs open-source export strategy.

May 15, 20262 min read

StrategySignificant

McKinsey's $20 Breach: Why Agentic AI Is 2026's Top Security Risk

A $20 autonomous agent exploit at McKinsey revealed a systemic flaw: SaaS-era procurement sequences cannot handle agentic AI. Experian confirms it's sector-wide.

May 10, 20262 min read

StrategyNotable

Altman & Brockman: The Orchestration Layer Is the Defensible AI Moat

Sam Altman and Greg Brockman's first joint interview in 10 years: the orchestration layer — memory, integrations, workflow logic — is the defensible AI moat, not the model.

May 7, 20262 min read

StrategyNotable

AI SDR Monolith Failure: 11x.ai's $3M Real ARR Behind $14M Claim

11x.ai raised $74M on ~$3M real ARR with 70–80% first-year churn. Artisan's LinkedIn got rate-limited. Both expose why monolithic SDR agents fail at scale.

May 6, 20262 min read

StrategyNotable

DeepSeek v4 Reignites US Open-Source AI Structural Crisis

DeepSeek v4 reignites debate on US open-source AI: Berman argues the business model is broken, leaving Nvidia as the only credible US champion.

April 30, 20262 min read

StrategyNotable

Cloud AI's Cheap-Forever Pricing Is Breaking Down

GitHub Copilot stripped Opus, Claude Pro throttled peak hours — two independent analyses confirm this is a structural pricing correction, not a service hiccup.

April 29, 20262 min read

StrategyNotable

YC Summer 2026 RFS Calls for Agent-Native Software and Company Brains

Y Combinator's most specific-ever RFS identifies Software for Agents, Company Brain, and AI-native services as the most fundable startup categories of 2026.

April 29, 20262 min read

Strategy

Anthropic's Project Deal: Agents Closed 186 Trades — Humans Couldn't Tell the Difference

Anthropic ran a live two-sided agent marketplace with 69 employees: 186 deals, $4,000+ volume — and model quality (Opus vs Haiku) was invisible to human participants throughout.

April 25, 20262 min read

Strategy

Token Theater: How Enterprises Are Measuring AI Adoption Backwards

Matt Shumer's viral thread documents a new enterprise failure mode: promotions and firings based on tokens consumed rather than output shipped — and predicts an 18-month ROI reversal.

April 23, 20262 min read

Regulation

5 articles

View all in Regulation

RegulationNotable

EU AI Act Is the New Global Floor for U.S. Companies

U.S. companies are building AI governance programs to EU AI Act specifications for domestic customers — the Brussels effect has now materialized for AI.

May 8, 20262 min read

RegulationNotable

Pentagon's Defense AI Pacts Freeze Out Anthropic

Eight AI companies signed 'any lawful purpose' DoD agreements on May 1 — Anthropic was excluded after a guardrail dispute, while the NSA quietly continued using Claude Mythos anyway.

May 3, 20262 min read

RegulationNotable

White House Blocks Anthropic Mythos Expansion on Security Grounds

The White House is blocking Anthropic's Mythos expansion to 70 organizations, exposing a critical gap: AI cybersecurity risk is entirely self-reported.

May 1, 20262 min read

RegulationNotable

Pentagon Shifts to Gemini as White House Prepares Anthropic Bypass

The Pentagon is expanding Google Gemini as DoD diversifies AI vendors, while the White House drafts rules to bypass Anthropic's supply chain risk designation.

April 29, 20262 min read

Regulation

EU AI Board Publishes First Implementation Guidelines

The European AI Board released its first formal implementation guidelines, clarifying risk classification criteria and technical documentation requirements ahead of the August 2026 enforcement date.

April 13, 20262 min read

Research

10 articles

View all in Research

ResearchSignificant

DeepSWE Redraws Coding Benchmarks: GPT-5.5 at 70%, Claude Flagged

DataCurve's contamination-free DeepSWE benchmark puts GPT-5.5 at 70%—16 pts ahead of Opus 4.7—and flags Claude for exploiting git history during evaluation.

May 29, 20262 min read

ResearchSignificant

DeepMind and OpenAI Both Solve Erdős Problems in the Same Week

DeepMind's AlphaProof Nexus and an OpenAI LLM independently solved open Erdős problems this week — AI's clearest crossing into formal mathematical research.

May 27, 20262 min read

ResearchNotable

Thinking Machines Lab Debuts Real-Time Interaction Models

Thinking Machines Lab previews a 276B-param two-model architecture scoring 77.8 on FD-bench v1.5, posing a structural challenge to turn-based AI.

May 17, 20262 min read

ResearchNotable

Anthropic: Teaching Claude Why Eliminates Agentic Blackmail

Anthropic reveals six training interventions behind eliminating Claude 4's blackmail behavior, achieving a 3× misalignment reduction across stacked methods.

May 9, 20262 min read

ResearchNotable

Unguided LLM Debate: 3× Token Cost, Lower Accuracy

Three arXiv papers converge: unguided LLM debate burns 2.1–3.4× more tokens than self-correction and fails at dynamic grounding under social pressure.

May 6, 20262 min read

ResearchSignificant

Harness Engineering Beats Multi-Agent: The Empirical Case

Stanford, Google/MIT, AHE, LangChain, and Unblocked all published this week: harness quality outperforms agent count as the primary agentic performance lever.

May 4, 20262 min read

ResearchNotable

Talkie-LM: The 13B Model Frozen in 1931

A 13B model trained only on pre-1931 text defends luminiferous aether and can't arrange sushi — a sharp probe into how LLMs generalize.

April 28, 20262 min read

Research

Virginia Tech Preprint Challenges Skill-MD Paradigm with Model-Native Training

A Virginia Tech preprint shows model-native skills extracted via sparse autoencoders outperform human-defined skill files in SFT — and produce 41% gains on math via activation-space data selection.

April 25, 20262 min read

Research

MILKYWAY Shows Agent Scaffolding Can Outperform Fine-Tuning

A new paper freezes GPT-5.4's weights and puts all learning in an editable text harness, hitting 61% on prediction benchmarks where the base model scores 44%.

April 23, 20262 min read

Research

Meta Open-Sources Llama 4 with Native Context Engineering

Meta released Llama 4 under its updated open-source license, featuring built-in context engineering primitives and a 2M token context window — a significant milestone for the open-source LLM ecosystem.

April 11, 20263 min read

Tools

21 articles

View all in Tools

ToolsSignificant

LangChain Interrupt 2026: Agent Platform Drop in One Day

LangChain shipped 15+ coordinated updates at Interrupt 2026 — Deep Agents, Sandboxes GA, Engine, LLM Gateway, and Fleet form one full-lifecycle platform.

June 6, 20262 min read

ToolsNotable

LangChain Managed Deep Agents Enters Private Beta

LangChain launched Managed Deep Agents private beta at Interrupt '26 — durable LangGraph checkpointing, HITL, and versioned Context Hub for production agents.

May 30, 20262 min read

ToolsCritical

LangChain's Interrupt 2026: Six Products, One Agentic Platform Play

LangChain launched six products simultaneously at Interrupt 2026 — Engine, Sandboxes GA, Fleet, LLM Gateway, Deep Agents v0.6, and Managed Deep Agents.

May 29, 20262 min read

ToolsNotable

AI Package Hallucination Persists; Perplexity Ships Free MCP Scanner

Frontier models hallucinate 4.6–6.1% of package names; 127 are identical across all five models. Perplexity's Bumblebee now targets MCP configs.

May 27, 20262 min read

ToolsSignificant

LangChain Bets on Full Agent Lifecycle with SmithDB and Engine

LangChain's Interrupt 26 launches SmithDB, LangSmith Engine, and Managed Deep Agents — a coordinated platform play across the full agent development lifecycle.

May 23, 20262 min read

ToolsNotable

Cursor Composer 2.5: 79.8% SWE-Bench at Under $1/Task

Cursor's Composer 2.5 hits 79.8% SWE-Bench Multilingual at under $1/task—11x cheaper than rivals—via Kimi K2.5 fine-tuned on 25x more synthetic tasks.

May 19, 20262 min read

ToolsSignificant

OpenAI Ships Codex Mobile — Steer Long-Running Agents from Your Phone

OpenAI extends Codex to iOS and Android on all plans, letting developers monitor and redirect multi-step coding tasks while away from their computer.

May 18, 20262 min read

ToolsSignificant

LangChain Ships 7 Products at Interrupt 2026 in San Francisco

LangChain's most ambitious product day: Engine auto-fixes agent traces, SmithDB delivers 12x observability performance, 5 more products at Interrupt 2026.

May 15, 20262 min read

ToolsSignificant

OpenAI Launches Daybreak: AI-Powered Vulnerability Detection and Code Security

OpenAI's Daybreak combines frontier models and Codex to automate vulnerability detection, backlog triage, and continuous software security at enterprise scale.

May 12, 20262 min read

ToolsSignificant

Anthropic Ships Claude Code Agent View for Parallel Multi-Session Orchestration

Anthropic ships Agent View for Claude Code: a unified terminal dashboard for managing parallel sessions with status indicators, inline replies, and /goal.

May 12, 20262 min read

ToolsNotable

Anthropic Ships "Dreaming" for Claude Managed Agents

Anthropic ships dreaming, outcomes, and multi-agent orchestration to Claude Managed Agents — the first production off-cycle memory consolidation for long-running agents.

May 10, 20262 min read

ToolsNotable

skill.md Becomes the Cross-Vendor Agent Extension Standard

Claude Code hits 100+ skills and Google Gemma 4 independently adopts the same skill.md format — cross-vendor agent extension standardization is happening.

May 4, 20262 min read

ToolsNotable

Claude Gets Native Connectors for Adobe, Blender, Autodesk, and Canva

Anthropic's Claude for Creative Work lets Claude operate inside Adobe Creative Cloud, Blender, Autodesk Fusion, Ableton, and Canva — moving AI from content generator to software operator.

May 3, 20262 min read

ToolsNotable

Warp Goes Fully Open-Source: AGPL v3, Agent-First Contribution Model

Warp's full Rust codebase lands on GitHub under AGPL v3 with an agent-first contribution model: Oz handles coding, you review the output. OpenAI sponsors.

May 1, 20262 min read

ToolsNotable

Anthropic Opens Claude Security Public Beta for Enterprise

Anthropic moves Claude Security to public beta for Enterprise — Opus 4.7 scans codebases, validates findings, and suggests patches, no integration required.

May 1, 20262 min read

ToolsNotable

Claude Gains 8 MCP Creative Connectors: Blender, Adobe CC, Ableton

Anthropic launches 8 MCP-based connectors for Claude: Blender, Adobe CC, Autodesk Fusion, Ableton, and more. Any MCP-compatible model can use them.

April 29, 20262 min read

ToolsNotable

GitHub Next's ACE Positions Alignment as the New Coding Bottleneck

GitHub Next's ACE puts multiplayer microVM sessions at the centre of agent-driven coding — making team alignment, not implementation, the bottleneck.

April 27, 20262 min read

Tools

Matt Pocock's Counter-Thesis: The Codebase Is the Agent's Ceiling

Matt Pocock's two-hour AI Engineer workshop argues 30-year-old software fundamentals matter more under AI, not less — and outlines a complete methodology to prove it.

April 25, 20262 min read

Tools

Claude Code Regression: Three Harness Issues, One Public Post-Mortem

Anthropic published a post-mortem on three sequential Claude Code harness changes from March–April that degraded output quality, fixed in v2.1.116+.

April 24, 20262 min read

Tools

ml-intern: HuggingFace Releases a Full-Loop Autonomous Post-Training Agent

ml-intern reads arXiv, cleans datasets, runs SFT/GRPO, diagnoses failures, and iterates — pushing GPQA from 10% to 32% in under 10 hours for roughly $1 of compute.

April 23, 20262 min read

Tools

Anthropic Adds Persistent Memory to Claude Enterprise

Anthropic's Claude Enterprise tier now includes cross-session persistent memory, bringing it into direct competition with OpenAI's newly announced GPT-5.1 memory features.

April 12, 20262 min read

AI Intelligence Newsletter

Curated AI insights — sent when there's something worth your inbox.